An enhanced technique for ground-based optical space debris extraction via RGA-based Neural Networks algorithm

• Main Authors: Torteeka, Peerapong, Peng-qi Gao, Ming Shen, Xiao-zhang Guo, Da-tao Yang, Huan-huan Yu, Wei-ping Zhou, Ming-guo Sun, You Zhao
• Format: Conference Proceeding
• Language: English
• Subjects: ground-based optical telescope; Heuristic algorithms; Neural networks; neural networks algorithm; Optical imaging; running Gaussian average; Space debris; Telescopes; Thyristors; Training
• ISSN: 2153-7003
• DOI: 10.1109/IGARSS.2016.7730591

A ground-based optical telescope equipped with CCD sensor has become an important tool for space debris monitoring in order to maintain a space debris catalogues. In this paper, we emphasized on an enhanced technique for a small and dimming space object extraction. Traditionally, the static background subtraction based on median image technique is widely used to extract the moving space object in astronomy field, because of its low computation cost. However, the main disadvantage of this technique is that it is not robustness with atmospheric effects, illumination changed and etc. Therefore, the proposed method, which is the combination of running Gaussian average and Neural Networks algorithm aim to enhance the traditional object extraction in order to precisely adapt the background model and also reduce the background noise. The performance of this proposed algorithm is evaluated by the experimental results with real background images, and showed that our proposed algorithm can achieve a high level of Signal-to-Clutter Ratio under a complex or dynamic-background.